Private AI that works from your documents — on hardware you own.

You decide, task by task. Everyday and sensitive work runs on the appliance in your home, stays there, and works even with no internet at all. When you choose to send a task to a cloud model like Claude or GPT, that request goes to the provider under your own account — and nothing else does.

Setup is guided from your phone. It arrives assembled and configured — no flashing, no monitor, no keyboard, no command line. Optional features like cloud models or messaging integrations require connecting your own accounts.

It's a capable personal AI assistant: answer questions, draft and summarize, organize your day, and run workflows you explicitly enable — with approved preferences and project context retained. It thinks on the device for everyday work and can call Claude, GPT, or Gemini through your account for the heaviest tasks. You can reach it from your phone or laptop.

Both kinds — that's the point. On the device, Hermes runs small, powerful open models, like Google's Gemma, that handle everyday work and keep private things private, even with no internet. When you want the most capable AI available, connect Claude, GPT, or Gemini with your own account, and Hermes hands the heavy thinking to them. You choose which brain gets which job.

You're buying the whole thing once — hardware, storage, and a private AI configured for you: $999 with 8GB of memory, or $1,999 with 16GB for larger local models. There's no required monthly fee because the assistant runs on hardware you own. Optional one-time services — like Life Upload, which loads your approved archive — are available whenever you want them.

You may request a return within 15 days of delivery; evaluation use is allowed. Return acceptance, return shipping, and any applicable restocking fee follow the Refund Policy.

Digital Twin Pro is built on NVIDIA Jetson — the family of small, efficient AI computers that powers robots, factory lines, and medical devices. Unlike an ordinary mini PC, a Jetson pairs its processor with a dedicated NVIDIA AI engine that shares one pool of memory, so it can run modern AI models around the clock while drawing about as much power as a light bulb. Silent, always on, and sized for a shelf — that's what lets Hermes live in your home instead of a datacenter. Each unit is assembled, configured, and supported in Miami, FL, USA.